Vertex z differences

Puzzle pieces regarding a possible east-west TPC displacement (in the drift direction, so either T0 or Drift Velocity, or both)


Run 7 Twist

From Andrew Rose's work on twist in Run 7 (2007):

FF (left) and RFF (right):


It appears that there was a ~1.3mm difference in Run 7 RFF data. I could not find his work for Run 9, but I did comment on it (see these Run 9 T0 differences between fields), indicating that the Twist calibrations (done after the T0 calibration) looked OK in z to within ~160 microns.


Run 9 T0 Calibrations

I see a Run 9 T0 differences between fields of about ~35 ns for all sectors between the two fields (FF needed larger T0s for the super-sector T0 calibrations than did RFF), with perhaps some smaller phi dependence (scale of ~±10ns) atop that. For a drift velocity of ~5.5 cm/μs, that corresponds to ~1.9 mm.


Run 9 z Residuals

I see a splitting of Residuals of membrane-crossing tracks in tracks crossing the central membrane in pp500 (RFF) data, also on the order of about 1.5 mm. Of note, the split is seen at all phi angles.


Run 9 z DCA

Justin Stevens sees z DCAs of global partners of primary tracks (I think) which are offset from the primary vertex in one direction for east side tracks, and in the other direction for west side tracks, adding up to roughly around 1.5 mm (see, for example, the mean of the histograms for Set A in figures 9 vs. 10 of More TPC distributions from L2W stream SL09g on TPC distributions in pp500 (RFF) data).


Run 9 BSMD Alignment

Willie Light sees a ~2 mm displacement between the east and west when matching TPC tracks to the BSMD in pp500 (RFF) data. See slide 5 of his BSMD Calibration Update: Alignment Study.

Jan Balewski documented more of Willie's observations in that same dataset BSMD - TPC allignment. While there may be a small z displacement at z=0 in the real data, there are additional displacements at locations corresponding to η of ±0.5, as well as some not-understood effects in the simulations.


Run 10 Twist

Maxim and Zubayer see a split of about 1.3 mm in RFF data. This is using T0s from the prompt hits.